Severe Sepsis and Septic Shock in 2014
Although sepsis resuscitation has been beaten to death by most ED residencies, with all of the new articles coming to surface and a paradigm shift in management I thought it might be a good time to complete a recap of my biased approach to septic shock in 2014. I welcome comments, suggestions and arguments as I know there will be some controversy in a few of the things we do.
65 y/o M presents to the emergency department with cough and fever for 2 days, and his wife noticed he was more sleepy than usual. His PMH is significant for COPD, HTN, DM II and his wife states she was recently sick with the “flu”.
VSS: 105, 85/55, 88% on 10L facemask, 39.6 (R), 26
My discussion will start with the first hour as I think these are all important things that can be done by the ED:
Obviously this is pounded into emergency physicians of fluids, antibiotics, and source control; here is maybe a few tips you haven’t heard about:
- Antibiotics:Give the most broad spectrum antibiotic FIRST in 99% of circumstances, like zosyn/cefepime or meropenem. These can be given as loading doses and rapid infusion over 30 minutes. Follow with your other MRSA coverage, which is usually vancomycin, but can be linezolid, dapto (non-pulmonary infections), tigecycline or ceftaroline depending upon the situation. Most nursing protocols won’t allow both antibiotics to be given together.
The dose of 1 g of vancomycin load in adult patients is almost always inadequate. Even if you don’t know the renal function you can adjust future doses of vancomycin to help stay within therapeutic range (15-20).
Subsequent dosing for me of these medications has changed to extended infusion at Stanford, and in some rarer circumstances continuous antibiotic infusion. The reason for this is as more MDRO infections arise it is important to have time above the MIC with time-dependent killers (beta-lactams) to help provide clinical and microbiologic cure.
Cephalosporins and penicillins have been shown that the amount of time that free or non-protein bound drug concentration exceed MIC is the best predictor of bacterial killing and microbiologic response for beta-lactams.
Traditionally we have given as much fluid as possible until the patient goes into pulmonary edema. I think it is reasonable to give a bolus of IVF in most patients of 1L who are hypotensive, and a second liter if still hypotensive.
Some principles of fluid resuscitation:
1. I like Plasmalyte, or normosol, but understand logistically it isn’t possible or financially acceptable. Also there are basically no studies assessing these particular fluids.
2. Otherwise 99% of the time you can use Lactated ringer, yes it has 4 meq of potassium, most of the time this will not make a huge difference in patients serum K. If need be alternate between LR and NS. Chloride restrictive strategy resulted in less renal failure and RRT in one trial, also in the Process trial higher fluid resus and positive fluid balance was associated with more RRT.
3. Avoid only using NS, More and more evidence is arising that it results in more renal dysfunction (we are seeing this a lot in patients with brain injury on 3% hypertonic saline). It also creates an acidosis that is very annoying to me :)
4. Colloids- the Albios study was a negative NEJM study on pushing the serum albumin up using 20% albumin (not as a resus fluid so not as applicable to ED) but a post-hoc subgroup analysis suggested 90 day mortality benefit in patients with septic shock. In the subgroup analysis of the SAFE trial albumin improved mortality in septic shock.
Therefofre I think albumin is an expensive blood product and is hyperosmolar (Na ~160ish) and it’s use is very controversial. I use it in dialysis patients, patients with cirrhosis and in patients who have already gotten crystalloid boluses and appear to still be volume responsive.
I don’t use starches (6S, CHEST and CRISTAL trial).
4. Try and use fluids that are warmed every liter of room temperature fluid you give will drop the patient’s temperature by 0.25 degrees Celsius.
5. Asses fluid responsiveness after you have loaded the patient with some fluids, using passive leg raising, IVC collapse, echocardiography with VTI, pulse pressure variation or whatever is available to you. Positive fluid balance is a strong predictor of death, results in pulmonary edema and more renal failure. If the nurse calls you and you are busy your response shouldn’t be ordering more fluids without reassessing after your initial load, especially if the patient is not intubated.
Fevers cause a whole myriad of problems like tachycardia, arrythmias, hypotension, hypoxia etc... but they are also protective in many animal studies by inducing production of protective heat shock proteins necessary for immune-mediated defense. I use tylenol for fever control 1g Q8, I do not use a cooling blanket or ibuprofen unless there is a really good reason ( i.e; the paitent has NMS or Malignant hyperthermia).
Blood Pressure Goals:
The big study SEPSISPAM by Asfar et al published in the NEJM 2014 compared 65-70 and 80-85. Basically no difference in mortality, but if you had baseline hypertension a higher MAP resulted in less renal replacement therapy (RRT). I shoot for 65-70, other people will stand up on their chairs and shout that this study proves what they suspected all along of patients with chronic hypertension needed higher perfusion pressures, go figure.
If you find that you have given fluids, and the patient is no longer fluid responsive or you notice that the blood pressure goes up for 10 minutes then back down it may be a problem of capillary leak and lack of vascular tone. Place a CVC above the diaphragm if possible at this point. I do not think the process trial should be interpreted as its now ok to avoid central lines, as that was not studied specifically.
- Vasopressors are in a way giving a fluid bolus. They decrease the unstressed volume in the venous splanchnic system where most of our blood volume is stored (basically I think of unstressed volume as the amount of volume you have to put into something before you can generate a pressure in it). Our bodies have evolved to deal with hypovolemia, not hypervolemia.
- Levophed (norepinephrine) - First line vasopressor for me and by the surviving sepsis guidelines. I use this until my dose starts to escalate above 5 mcg/min.
- Vasopressin-I use this in patients who have levophed requirements above 5 mcg/min. My non-titrating dose of vaso is 0.04 U/min. Vasopressin has a different mechanism of action than all other vasopressors (that work as alpha agonists). Vasopressin inhibits potassium channels that result in production of nitric oxide a potent vasodilator and an inflammatory mediator in sepsis. This makes vasopressin both anti-inflammatory and a potent vasoconstrictor. The VASST trial by Jim Russel’s group in Vancouver showed it improved mortality in patients who had lower doses of vasopressors (just hypothesis generating). It also lowers the dose required of norepinephrine. When combined with steroids appears to lower requirements for vasopressors all together (see below).
- Epinephrine- I use epinephrine if the patient has myocardial dysfunction based on bedside echo. I also reevaluate in real-time, VTI and EF after starting epinephrine.
A few things to know about epi:
1. It will make you blood sugar higher and
2. It will make your lactate go up.
At does of 50 nanos/kg/min or 0.05 mcg/kg/min or less epi is considered mostly inotrope dosing. I therefore try and use it to help augment CO early in septic shock since most patients are peripherally vasodilated, later in there course is when they become vasoconstriced and dobutamine may be a better option.
- Phenylephrine: I don’t use phenylephrine except for unstable Afib to try and cause some bradycardia after bolus dosing and post-procedural hypotension from vasodilating agents like propofol.
- Milrinone- Cardiac ICU in patients with pulmonary artery catheters or select cases of patients with pulmonary hypertension. Probably not gonna get used in ED.
- Levosimendan – Don’t have any experience with the drug because it’s not available here. Probably something to be on the lookout for in near future.
I start them early and do not wait for a day of shock to initiate steroids. If you require two vasopressors (that is vaso @ 0.04 U/min and norepi at 5 mcg/min) you get hydrocortionse 200mg 24-hour infusion.
An infusion is hard to get in the ED but 50mg IVP is pretty easy to find, and so 50 Q 6 is reasonable until they get upstairs. The physiology of a continuous infusion seems to make sense, and there are a few studies showing better glucose control and maintenance of cortisol levels.
Also I don’t usually check random cortisol levels or do a stim test, mostly because it takes time to come back, and because no one really knows the definition of adrenal insufficiency in the ICU. There are a ton of factors that go into a random cortisol level that I won’t get into here.
Hydrocortisone is dirt-cheap and the side effects of hyperglycemia are not as significant as you would expect. The superinfection rates are about the same between those getting low dose steroids and placebo (Sligli et al. Safety and Efficacy of Corticosteroids for the Treatment of Septic Shock: A systematic Review and Meta-Analysis). Also we know that steroids will reduce the incidence of vasopressor dependency. In my opinion (as well as many other ICU docs) the CORTICUS study waited WAY too long to enroll someone, there was already a general consensus in Europe that steroids worked (based on the Annane paper published earlier) and the patients were less critically ill than in the original paper. Therefore, in a select subset of patients (2 vasopressors; even if low dose), we start steroids, and RAPDILY taper in ICU as the vasopressors start to come down.
Lactate-I order two to start, one at time 0, and one at To + 4 hours. This way no matter what that lab gets drawn 4 hours after the original in order to make sure we are making progress at the micro circulatory level.
Also if I have high levels of lactic acidosis not responding to most treatments (and no other apparent cause) I supplement thiamine, in case there is a reason they could be thiamine deficient, since this medication is cheap, and basically inert.
A caveat: lactate is not sensitive or specific for sepsis, it is a marker of poorer prognosis in patients in shock of any cause (post-arrest, septic, cardiogenic etc). If you get an elevated lactate do not assume the patient is septic and call the ICU fellow, do some digging first.
SvO2- probably don’t have to routinely measure this. If the patient is in shock needs vasopressors and has a central line above the diaphragm I will follow SvO2s as I am titrating vasopressors and Inotropes. I don’t transfuse PRBC’s or use dobutamine to push this “number” up.
I check a CVP, so I know what it is, and then look at it in conjunction with 700 other things. I do not bolus fluids to get a certain CVP number. The CVP is based on multiple things and does not predict fluid responsiveness nor does a static CVP give me much information. I would caution that if the CVP is high there is something wrong with the right side of the heart and it should prompt an echocardiogram. It could be LV failure, pulmonary hypertension, PE, RV dysfunction due to ischemia etc, but a CVP > 16 should give you pause, your cvp lying down is probably 0.
Setting up the Vent:
This is a whole article in itself, but I leave you with the basics:
Your hospital needs to develop a lung protective protocol that is RT driven, to take you off the hook. We call this protocol LPV (lung protective ventilation). Basically the RT should start with 8cc/kg IBW for TV and a RR to match their current Minute ventilation. You can then titrate down on TV to 6cc/kg, as tolerated to keep plateau pressures < 30. You should titrate up on PEEP to keep sats up not the FiO2. If your patient is on 70% fiO2 and PEEP of 5, you probably need to rethink that strategy. If your patient is hypoxic do not be afraid of PEEP! An RT driven protocol frees you up and allows them to make adjustments based on ARDSnet, which I know you don't want to be doing.
The FiO2 should be reduced to keep sats >88%, you don’t need an ABG, it should be weaned rapidly.
The pH can be tolerated down to as low as 7.2. Keep in mind MOST patients aren’t this sick to need lower pH’s and patients HATE having an elevated PCO2. It is terribly uncomfortable and often requires high doses of opiates and sedatives. Most patients do fine with AC/VC with low TV and rates between 16-20.
Our sedation/analgesia package right after intubation if they are hemodynamically stable is:
-1-2mg of dilaudid IVP after securing airway and checking BP, then
1 mg/hr drip of dilaudid for avg adult male (less for granny and more for chronic opiate users), we have abandoned fentanyl drips due to the context sensitive half-life. Others use intermittent pushes of opiates, in the ED that is way too much work, they aren’t getting extubated in an hour or so, but if you are brave please go ahead.
I rarely use benzodiazepines, PERIOD. They increase ICU related delirium, drips result in prolonged wake-up times and if not vigilant about sedation vacations these can hang around for a good deal of time. Once the patient has been on them for a while they can go into benzo withdrawl, another annoying thing to deal with. Basically STOP USING BENZO’s, enough literature is coming out to suggest benzo’s are dangerous. We have gone as far as not using benzo’s in alcohol withdrawl. We use clonidine, dexmedetomidine, gabapentin, and Depakote. (this is a whole different lecture again).
Dexmedetomidine will go generic this winter. Also even though the price tag gives administrators and pharmacists a heart attack, evidence suggests that on the back end dex provides cost savings with shorter ICU LOS, and reduced hospital costs.
It works in about 50-70% of patients to keep them calm on the vent but comfortable enough to interact. The dose is 0.2- 1.2 mcg/kg/hr. I don’t bolus and the major adverse events are bradycardia. Get familiar with this med as it will become a great ED med for obtaining CT’s, conscious sedation for anxiety provoking procedures like LP’s in the near future.
Propofol is excellent for its rapid on and off action but I avoid it in patients with hypotension, for obvious reasons. Avoid doses above 80 mcg/kg/min, as these are associated with propofol infusion syndrome(PRIS) after 24-48 hours. PRIS is devastating and took the life of a pregnant 35 y/o with eclampsia due to high doses at an OSH as we were the recieving hospital of the transfer for ECMO.
Ketamine works but I usually use it as an adjunctive medication to one of the above (ketamine/dex/dilaudid).
I tried to be somewhat breif and still controversial and maybe give a few pointers of thing you had forgotten, let me know what your thoughts are.